Fast and Efficient Multistage Pedestrian Detection Method Using Computer Vision with Applications to 5G/6G Communications

The paper presents a fast an efficient pedestrian detection method using computer vision. The method is based on a multistage approach combining motion or change detectors with heuristics and small size convolutional neural networks (CNN). The method is faster than direct application of larger CNNs but produces comparable results. Possible application of the method in 5G/6G millimeter wave communication systems is discussed.